针对桥梁应变监测信号存在多源噪声,以及现有降噪法难以准确选取有效固有模态函数(Intrinsic Mode Function,IMF)分量的问题,提出一种有效的多滤波降噪法.首先,采用自适应噪声抵消器对含噪信号进行预处理,以滤除低频噪声,再对其进行自适应噪声完备集合经验模态分解(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise,CEEMDAN)得到IMF分量,并利用频谱分析-自相关函数双重判断准则选取IMF分量;随后,对其进行奇异值分解,利用奇异值差分谱确定各分量有效阶数;最后,将各有效IMF分量的有效阶重构得到去噪信号.通过模拟试验验证上述方法的合理性,并将其应用于桥梁应变监测信号处理.结果表明:采用上述双重判断准则选取有效IMF分量具有较好效果,且提出的多滤波降噪法在桥梁应变监测信号处理中具有显著的优越性.
Bridge Strain Monitoring Signal Processing Based on Multi-filter Denoising Method
In view of the issue that multi-source noise exists in bridge structural strain monitoring signals,and the prob-lem that effective intrinsic mode function(IMF)components are difficult to be selected accurately with existing denoising methods,a multi-filter denoising method was proposed.Firstly,the noisy signal was preprocessed by an adaptive noise can-celer to filter out low-frequency noise.Then,the preprocessed signal was decomposed by complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)into several IMF components.Furthermore,a double judgment crite-rion of spectrum analysis and autocorrelation function was employed to select the effective IMF components.Subsequently,the selected effective IMF components were decomposed into singular values,and their effective orders were determined by singular value difference spectrum.Finally,the denoised signal was obtained by superimposing the effective IMF compo-nents of the effective orders.The rationality and feasibility of the proposed multi-filter denoising method was verified by simulation tests.This method was utilized in processing the bridge structural strain monitoring signals.The results indicate that the selection of effective IMF components using the double judgment criterion has better effect,and the proposed denois-ing method has significant advantages in bridge structural strain monitoring signal processing.
vibration and wavebridge stainmulti-filter denoising methoddouble judgment criterionintrinsic mode functionCEEMDANsingular value decomposition
万方数据
维普
